The Penn Discourse Treebank

The Penn Discourse Treebank (PDTB) is a corpus developed at the University of Pennsylvania (UPenn). The PDTB is, to date, the largest annotation effort at the discourse level, providing annotations of explicit and implicit discourse connectives. It provides annotations of the argument structure, attribution and semantics of discourse connectives. PDTB annotations relate a discourse connective with its two arguments.

You can visit the official site of the PDTB by following this link.

When I was visiting University of Pennsylvania (from February to July 2007 and later From June to July 2009) I started working with the PDTB research group, in particular with Eleni Miltsakaki . I helped the writing of the PDTB 2.0 annotation manual and the sense annotation in the release 2.0 of the corpus. The sense annotation done in the PDTB is described in the manual as well as in (Miltsakaki et al., 2008).

In (Robaldo and Miltsakaki, 2014), we conducted an empirical study on 1000 occurrences of concessive relations taken from the PDTB. The analysis led to the identification of four possible sources of Expectation. Then, a uniform formalization of three of those sources in the reification-based logic defined by prof. Jerry R. Hobbs is proposed. This was my last contribution to the project.